Session Number

22

Description

The information explosion of the last century has had profound effects upon the entire society: The number of publications has increased enormously, and important consequences have been the development of bibliographic tools, of libraries and an improved cooperation between libraries. Computer technology has made a storage of bibliographic information in big, international data bases possible, and complex networks are now being established. As bibliographies and catalogues in their paper-ver si ons thus are disappearing, other sorts of literature are going to be replaced by on-line accessible data bases, and telephone directories, sales catalogues etc. are soon to disappear in their traditional form. The question is whether the "literature proper" - books and articles - are going to be computerized too. Already in 1969 F. W. Lancaster predicted that a complete change to a paperless society should be completed in 1990, and as late as 1979 Chris. Evans advocated exactly the same point-of-view. Other new media such as microforms, video-disks, teletext, and videotex are part of the ongoing diversification as far as information storage and transrnission is concern ed. At the EURIM 4 Conference, Brussels 1980, a half-way status was made, and John W. Senders reported of an experiment at producing an electronic periodical without any paper version at all. In the U.S. other experiments are recently being performed, aiming at storing the text of entire books (The Bible a.o.) by means of microprocessors to be used in connection with a reading plate device on which the text can be blown up. The explosive growth of literature has created a problem of abundance which cannot be solved by means of data bases and computerization alone: First the data bases usually give far too much information - too many references compared with the needs and capacities of the users. Second the very lack of redundancy of the output is at the same time a strength and a weakness seen in relation to the actual search situations: Apart from very specific subject inquiries performed by research workers, technicians, and other experts, the great majority of typical document search activities are best characterized by the facts that the users do not at all want information about all existing literature of a given subject. On the contrary: Users want to get one or a few, well selected articles relevant in the given context. It is, therefore, highly important, that user studies dealing with such badly defined search situations are being performed. They are, indeed, extremely common in public libraries with their mixed, and entirely heterogeneous clientèle; user studies of this kind could therefore reasonably be initiated in this type of libraries (T. Johansen, P. Ingwersen and P. Timmermann, 1980) (Annelise Mark Pejtersen (1980). Citation analyses show unambiguously that about half the articles published are never cited - that means: are never used by later research wor kers within the same subject field. Results of user studies - of any type - should be combined with comparative investigations of a number of information retrieval systerns (i.e.: bibliographies, indexes, abstract services, catalogues, data bases etc.); the Cranfield experiments indicate quite clear, that redundance-free or loss-free systems do not exist in the real world. And: Recall and precision are related factors.

Share

COinS
 
Apr 16th, 12:00 AM

Dilemmas of Document Communication

The information explosion of the last century has had profound effects upon the entire society: The number of publications has increased enormously, and important consequences have been the development of bibliographic tools, of libraries and an improved cooperation between libraries. Computer technology has made a storage of bibliographic information in big, international data bases possible, and complex networks are now being established. As bibliographies and catalogues in their paper-ver si ons thus are disappearing, other sorts of literature are going to be replaced by on-line accessible data bases, and telephone directories, sales catalogues etc. are soon to disappear in their traditional form. The question is whether the "literature proper" - books and articles - are going to be computerized too. Already in 1969 F. W. Lancaster predicted that a complete change to a paperless society should be completed in 1990, and as late as 1979 Chris. Evans advocated exactly the same point-of-view. Other new media such as microforms, video-disks, teletext, and videotex are part of the ongoing diversification as far as information storage and transrnission is concern ed. At the EURIM 4 Conference, Brussels 1980, a half-way status was made, and John W. Senders reported of an experiment at producing an electronic periodical without any paper version at all. In the U.S. other experiments are recently being performed, aiming at storing the text of entire books (The Bible a.o.) by means of microprocessors to be used in connection with a reading plate device on which the text can be blown up. The explosive growth of literature has created a problem of abundance which cannot be solved by means of data bases and computerization alone: First the data bases usually give far too much information - too many references compared with the needs and capacities of the users. Second the very lack of redundancy of the output is at the same time a strength and a weakness seen in relation to the actual search situations: Apart from very specific subject inquiries performed by research workers, technicians, and other experts, the great majority of typical document search activities are best characterized by the facts that the users do not at all want information about all existing literature of a given subject. On the contrary: Users want to get one or a few, well selected articles relevant in the given context. It is, therefore, highly important, that user studies dealing with such badly defined search situations are being performed. They are, indeed, extremely common in public libraries with their mixed, and entirely heterogeneous clientèle; user studies of this kind could therefore reasonably be initiated in this type of libraries (T. Johansen, P. Ingwersen and P. Timmermann, 1980) (Annelise Mark Pejtersen (1980). Citation analyses show unambiguously that about half the articles published are never cited - that means: are never used by later research wor kers within the same subject field. Results of user studies - of any type - should be combined with comparative investigations of a number of information retrieval systerns (i.e.: bibliographies, indexes, abstract services, catalogues, data bases etc.); the Cranfield experiments indicate quite clear, that redundance-free or loss-free systems do not exist in the real world. And: Recall and precision are related factors.